The vignetting effect is a widespread phenomenon in the process of infrared imaging. Because the output brightness or contrast of the infrared imaging system is low, the impact of vignetting effect on the performance of the system is very serious. Therefore, it is particularly important to correct the vignetting effect. The cause of the generation of the vignetting effect was analyzed. Firstly, a fast correction method was proposed. The difference between scene was used to get background information. The background gray level distribution was approximated by the two-dimensional Gauss distribution, and the correction factor was obtained. Then the background extraction method was improved by using the variance information, and the method was applied to a more complex scene. The concept of the sum of pixel difference was presented. After correction, the sum of pixel difference decreaseal to 1/5-1/2 of the sum of pixel difference before correction. The correction effect of the two methods is better without using a reference scene of uniform gray distribution.